November 5, 2025 13:00 CEST
Live on Microsoft Teams.
On 5 November at 13:00 CEST, the ESA Φ-Lab Collaborative Innovation Network will host a new Φ-talk. Details are below.
Meet the speakers
Dan Guest is a particle physicist based at CERN, affiliated with the Humboldt University of Berlin. His research focuses on developing innovative and physics-driven methods for particle physics at the ATLAS experiment. Notably, he drove the effort to use modern ML for flavour-tagging, a particle identification step crucial for measuring the Higgs boson. Currently, he serves as the convener of the LHC Machine Learning Working Group representing the ATLAS experiment, where he continues to lead innovative efforts in ML for physics simulation, particle reconstruction & identification and physics searches.
William Fawcett is a long-time machine learning enthusiast. Spending much of his career with the ATLAS experiment, he worked on searches for new hypothetical particles beyond the standard model of particle physics. He is now Lead Research Engineer at Trillium Technologies, who partners with ESA and NASA to deliver the "Frontier Development Lab", seeking to bring cutting-edge AI research to problems in the space sector. He maintains an academic connection with the University of Cambridge as Director of Studies in Physical Natural sciences at Homerton College.
Recently graduated from the University of Oxford with a PhD in particle physics, Maggie Chen has been tackling a wide range challenges for the ATLAS experiment using machine learning, including particle identification and searches for rare physics processes producing multiple Higgs bosons. Now, she continues to expand the use of ML to other fields. In partnership with Frontier Development Lab and ESA, she is developing ML-driven solutions for methane detection that can be deployed onboard spacecrafts.
Talk abstract
Accelerating protons to near the speed of light deep underground at CERN, the Large Hadron Collider (LHC) is the most powerful particle accelerator on Earth, attempting to answer some of the most fundamental questions about the Universe. At the ATLAS detector, proton bunches are brought to collide millions of times per second, such that particles like the Higgs boson and potential dark matter candidates can be created and investigated from the petabytes of data collected each year. Under an environment dense with particles inside a detector of extremely complex design, searching for these interesting particles is often akin to looking for a needle in a haystack. Scientists at the LHC have been developing ML methods to tackle a broad range of challenges. In this talk, three researchers will present some examples and their unique perspectives on ML applications in particle physics. The topics include particle reconstruction & identification using transformers, using tensor-attention networks to build automated “structure-finding” models and using dataset-wide graph neural networks to search for new particles. By drawing parallels between the challenges in particle physics and earth observation at ESA, scintillating discussions and valuable knowledge exchange are to be the highlight of this seminar.